Ma. Stone et al., EFFECTS OF FAST-ACTING HIGH-FREQUENCY COMPRESSION ON THE INTELLIGIBILITY OF SPEECH IN STEADY AND FLUCTUATING BACKGROUND SOUNDS, British journal of audiology, 31(4), 1997, pp. 257-273
This study examines whether speech intelligibility in background sound
s can be improved for persons with loudness recruitment by the use of
fast-acting compression applied at high frequencies, when the overall
level of the sounds is held constant by means of a slow-acting automat
ic gain control (AGC) system and when appropriate frequency-response s
haping is applied. Two types of fast-acting compression were used in t
he high-frequency channel of a two-channel system:a compression limite
r with a LC:1 compression ratio and with a compression threshold about
9 dB below the peal: level of the signal in the high-frequency channe
l; and a wide dynamic range compressor with a 2:1 compression ratio an
d with the compression threshold about 24 dB below the peak level of t
he signal in the high-frequency channel. A condition with linear proce
ssing in the high-frequency channel was also used. Speech reception th
resholds (SRTs) were measured for two background sounds: a steady spee
ch-shaped noise and a single male talker. All subjects had moderate-to
-severe sensorineural hearing loss. Three different types of speech ma
terial were used: the adaptive sentence lists (ASL), the Bamford-Kowal
-Bench (BKB) sentence lists and the Boothroyd word lists. For the stea
dy background noise, the compression generally led to poorer performan
ce than for the linear condition, although the deleterious effect was
only significant for the 10:1 compression ratio. For the background of
a single talker, the compression had no significant effect except for
the ASL sentences, where the 10:1 compression gave significantly bett
er performance than the linear condition. Overall, the results did not
show any clear benefits of the fast-acting compression, possibly beca
use the slow-acting AGC allowed the use of gains in the linear conditi
on that were markedly higher than would normally be used with linear h
earing aids.